Andy, That seems like a reasonable schedule. I have 2 VPS and have done the changes. I did hit one problem, though it was easily sorted. My first VPS went like a dream, took just a couple of minutes to do then I did a shutdown and reboot to make everything permanent. The second one, everything went smoothly, both addresses working OK. I did the shutdown and reboot but then it came back with the message that it could not read /etc/network/interfaces (I'm using Debian) I checked ownership and permissions, these were all OK. But the VPS was unreachable from the network. The solution, before making any changes, I had downloaded and printed out a copy of the file. I renamed /etc/network/interfaces /etc/network/backinterfaces then using nano (never did like vi) typed it in manually. Shutdown and booted and all is working. I have no idea what went wrong with the interfaces file first time round, but if anyone else has a similar problem, that's the answer! Keith On 5 November 2011 03:57, Andy Smith <andy(a)bitfolk.com> wrote: -- Keith Williams If the only tool you have is a hammer, you tend to see every problem as a nail. - Abraham Maslow ทำดีได้ดี ทำชั่วไดชั่ว

Thanks for the reminder, Andy. I am dimly aware of the instructions you listed on a bitfolk webpage-wiki (for which thanks), and also of the followup emails on 'rndc' and other commands necessary to have bind apply these changes.  Is everything now updated on that bitfolk webpage, so that people can check that they have indeed taken all the right steps? --- On Tue, 1/31/12, Andy Smith <andy(a)bitfolk.com> wrote: From: Andy Smith <andy(a)bitfolk.com> Subject: Re: [bitfolk] IMPORTANT: You need to renumber the IP address(es) of your BitFolk VPS To: users(a)lists.bitfolk.com Received: Tuesday, January 31, 2012, 8:38 PM Hello, On Sat, Nov 05, 2011 at 03:57:40AM +0000, Andy Smith wrote: I've been sniffing packet headers and sending out warnings for the past couple of weeks, and in the cases where people have replied to ask for more info it seems like it has mostly ended up now being traffic from scans and other unwanted garbage. That's a good thing, but I can't easily tell what is garbage and what isn't, both from a privacy point of view of looking at your network traffic and from a workload point of view. So, I would recommend that if you have any of the deprecated IP addresses still active that you disable them as soon as possible. That will stop you receiving any traffic to those IPs => will stop you receiving any scans to those IPs => will stop you receiving warning emails from me. It will also provide a good final check as to whether you really have caught everything. The alternative to shutting them off when you like is to keep receiving (possibly irrelevant) warnings up until the point when they are shut off at some indeterminate time next week. I think shutting them off at a time that suits yourself would be better! Cheers, Andy -- http://bitfolk.com/ -- No-nonsense VPS hosting -----Inline Attachment Follows----- _______________________________________________ users mailing list users(a)lists.bitfolk.com https://lists.bitfolk.com/mailman/listinfo/users

Yes, Andy, thanks for the pointer. I would have liked to update '6 Reconfiguring services' to point to a new page in order to summarise the 'bind' discoveries of Michael Corliss*.  Unfortunately I am unable because I lack information about my 'login name' and 'password', if ever I had one. *I am surprised his discoveries hadn't been included already.  Does nobody besides us use bind?? --- On Thu, 2/2/12, Andy Smith <andy(a)bitfolk.com> wrote: From: Andy Smith <andy(a)bitfolk.com> Subject: Re: [bitfolk] IMPORTANT: You need to renumber the IP address(es) of your BitFolk VPS To: users(a)lists.bitfolk.com Received: Thursday, February 2, 2012, 1:43 PM Hi Max, Do you mean:     https://tools.bitfolk.com/wiki/Renumbering_for_customers ? If you have removed the deprecated IP address(es) from your VPS and it still works, you have taken the right steps. :) Cheers, Andy On Thu, Feb 02, 2012 at 12:52:37AM -0800, Max B wrote: -- http://bitfolk.com/ -- No-nonsense VPS hosting -----Inline Attachment Follows----- _______________________________________________ users mailing list users(a)lists.bitfolk.com https://lists.bitfolk.com/mailman/listinfo/users

Andy, your view of 'renumbering for clients' and mine differ.  The way I look at it, any issues that force me to deal this change are directly relevant to 'renumbering for clients'.   I am surprised you appear to disagree. Am I a client or not?  This is a serious question, although it may appear to be flippant. If Michael's discovery of a 'rndc reload' command does not fall under '6 Reconfiguring services', I don't know what does.  Apparently, neither does anyone else, given the lack of contributions to this section. 'rndc reload' is a misnomer; traditionally, under system V unix naming conventions, 'bind' would restarted by an /etc/init.d script, either 'bind' or a catchall like 'networking'.  'rndc reload' is odd enough to deserve a comment on 'renumbering for clients'. Cheers, Max --- On Mon, 2/6/12, Andy Smith <andy(a)bitfolk.com> wrote: From: Andy Smith <andy(a)bitfolk.com> Subject: Re: [bitfolk] IMPORTANT: You need to renumber the IP address(es) of your BitFolk VPS To: users(a)lists.bitfolk.com Received: Monday, February 6, 2012, 12:24 AM Hi Max, I'm afraid I don't really know what you mean about Michael's bind discoveries. All I can recall from that thread is that he found that he had forgotten to reload his nameserver after making changes to a zone file. That isn't really related to renumbering a VPS so I don't think it belongs on the renumbering article. Cheers, Andy On Sun, Feb 05, 2012 at 02:16:53PM -0800, Max B wrote: -- http://bitfolk.com/ -- No-nonsense VPS hosting -----Inline Attachment Follows----- _______________________________________________ users mailing list users(a)lists.bitfolk.com https://lists.bitfolk.com/mailman/listinfo/users

Thanks, Ewan, for the two nuggets... FWIW, I have yet to reboot my machine.  I suppose it's fear: if it ain't broke, don't fix it.  ;)  uptime now 344 days since last (bitfolk-wide?) reboot. Cheers! Max --- On Mon, 2/6/12, Ewan Leith <ewan(a)nutmegdata.co.uk> wrote: From: Ewan Leith <ewan(a)nutmegdata.co.uk> Subject: Re: [bitfolk] IMPORTANT: You need to renumber the IP address(es) of your BitFolk VPS To: "Max B" <txtmax(a)yahoo.ca> Cc: users(a)lists.bitfolk.com, "Andy Smith" <andy(a)bitfolk.com> Received: Monday, February 6, 2012, 9:44 AM Hi Max, I've updated the wiki now to mention the "rndc reload" command and updating your zone files if you use bind - the wiki login is just your usual bitfolk username. https://tools.bitfolk.com/wiki/Renumbering_for_customers#Reconfiguring_serv…  I suspect most people will of just rebooted the box (or restarted all the bind services) if they were having issues with DNS, which would have forced a read of the updated zone file. Ewan On Mon, Feb 6, 2012 at 9:12 AM, Max B <txtmax(a)yahoo.ca> wrote: Andy, your view of 'renumbering for clients' and mine differ.  The way I look at it, any issues that force me to deal this change are directly relevant to 'renumbering for clients'.   I am surprised you appear to disagree. Am I a client or not?  This is a serious question, although it may appear to be flippant. If Michael's discovery of a 'rndc reload' command does not fall under '6 Reconfiguring services', I don't know what does.  Apparently, neither does anyone else, given the lack of contributions to this section. 'rndc reload' is a misnomer; traditionally, under system V unix naming conventions, 'bind' would restarted by an /etc/init.d script, either 'bind' or a catchall like 'networking'.  'rndc reload' is odd enough to deserve a comment on 'renumbering for clients'. Cheers, Max --- On Mon, 2/6/12, Andy Smith <andy(a)bitfolk.com> wrote: From: Andy Smith <andy(a)bitfolk.com> Subject: Re: [bitfolk] IMPORTANT: You need to renumber the IP address(es) of your BitFolk VPS To: users(a)lists.bitfolk.com Received: Monday, February 6, 2012, 12:24 AM Hi Max, I'm afraid I don't really know what you mean about Michael's bind discoveries. All I can recall from that thread is that he found that he had forgotten to reload his nameserver after making changes to a zone file. That isn't really related to renumbering a VPS so I don't think it belongs on the renumbering article. Cheers, Andy On Sun, Feb 05, 2012 at 02:16:53PM -0800, Max B wrote: pointer. -- http://bitfolk.com/ -- No-nonsense VPS hosting -----Inline Attachment Follows----- _______________________________________________ users mailing list users(a)lists.bitfolk.com https://lists.bitfolk.com/mailman/listinfo/users _______________________________________________ users mailing list users(a)lists.bitfolk.com https://lists.bitfolk.com/mailman/listinfo/users -----Inline Attachment Follows----- _______________________________________________ users mailing list users(a)lists.bitfolk.com https://lists.bitfolk.com/mailman/listinfo/users

On 06/02/2012 09:12, Max B wrote: Hi Max, I'd be inclined to view the information in that wiki page as being a courtesy to elucidate the basics of the numbering change, as far as it *directly relates* to Bitfolk's networking infrastructure, rather than setting out details for all possible services and other dependencies. e.g. there's no mention there on updating your domain registration if you self-host your DNS, no details on how to edit zone files, or update httpd configs if you had explicit IPs in there, update SSL certificates, VPN configurations etc. For bind9, a reload might be invoked in various ways depending on the distro and personal taste (I still use "invoke-rc.d bind9 reload") and documenting that level of detail is to my mind well out of scope for a renumbering guide. You might note that the front page of the wiki states "This wiki is here for BitFolk customers to contribute their own content". I can imagine many ISPs giving you barely a one-liner "On <date> your new IP address will be <x>, gateway <y>"; customers savvy enough to want to run their own servers might reasonably be expected to have the competencies to effect all the changes required. This in turn might suggest a reason for the paucity of contributions to that wiki section. Perhaps you're conflating the implications of "renumbering for clients" with those for "renumbering clients"? -jmb

Hi Max, On Mon, Feb 06, 2012 at 01:12:43AM -0800, Max B wrote: I feel like you would end up having to list many many different possible applications that may need to be reconfigured, plus all the myriad ways in which one could possibly fail to do so, none of which is particularly related to renumbering a VPS. That said, it is a wiki, go ahead and write what you like. Where is your wiki edit? This is a serious question, although it may appear to be flippant. There is no specific mention of Apache either. By your logic, Apache should be mentioned, with a note that says "don't forget to really reload it after you've changed the config". If you run rtorrent to distribute your software, and have it bound to a specific address, don't forget to change that when you renumber otherwise it will just sit there appearing to run but not actually doing anything. And don't forget to reload it after you change the config. We could be here for weeks listing these things. But notes for specific software are not going to hurt anyone, so in my opinion go ahead if you feel it's worth mentioning. "rndc" is a part of BIND; it's a client that can tell it to do many things, from reload the config to reload an individual zone. Yes the same effect can be achieved from restarting the process with the init script, but this would for example stop the process before starting it again and possibly failing due to configuration error. On Debian, "invoke-rc.d bind reload"¹ just calls "rndc reload". I've never heard of anyone saying "rndc" is not the correct way to administer a BIND process before, given that is what it's for. If rndc didn't work, I'd consider the BIND configuration broken. Cheers, Andy ¹ The way you are supposed to call init scripts in /etc/init.d/ on Debian

On Sat, 2011-11-05 at 10:48 -0700, Max B wrote: If I understand you correctly, then all you need to change on the server are the zone files and the IP address BIND is listening on for external requests (you'll want it to listen to both the old and the new address for now). Depending on how you DNS is set up, you may also need to update your glue record with GoDaddy. M

Hi Michael, On Sun, Nov 06, 2011 at 02:12:02PM +0000, Michael Watkin wrote: […] Brilliant, thanks. I think it's the addition of "dev eth0" which made the difference here. I'll update the page since it's not going to cause any harm. Sandy, can you confirm this (with the "metric 100" part) works for you? Cheers, Andy -- http://bitfolk.com/ -- No-nonsense VPS hosting

On 6/11/2011 4:57 ??, S P wrote: Does it not look to you like this VPS had no default route specified? Maybe this is why the command did not work. -- ?? ????????, ?. ???????? Elementality ????????? ???????????? corfiot(a)elementality.biz

Hi, On Sun, Nov 06, 2011 at 08:28:41PM +0200, G. Miliotis wrote: Was it definitely the "metric 100" that you needed as opposed to just the "dev eth0"? Cheers, Andy -- http://bitfolk.com/ -- No-nonsense VPS hosting

On Sun, Nov 06, 2011 at 09:52:56PM +0200, G. Miliotis wrote: OK, thanks for investigating! I've updated the wiki article: https://tools.bitfolk.com/w/index.php?title=Renumbering_for_customers&a… Cheers, Andy -- http://bitfolk.com/ -- No-nonsense VPS hosting

Hi, In what was did it not work? What's wrong with that? ----- Metric The ’distance’ to the target (usually counted in hops). It is not used by recent kernels, but may be needed by routing daemons. ----- You should only ever have to set the metric manually in order to make a route be considered expensive. It's only really an important value when there's a choice of routes. Regards, @ndy -- andyjpb(a)ashurst.eu.org http://www.ashurst.eu.org/ 0x7EBA75FF

On 6/11/2011 10:38 μμ, Andy Bennett wrote: The command failed. It gave the same error we have been discussing. -G

On 6/11/2011 10:49 μμ, Andy Bennett wrote: The commands I showed were not executed in order. I deleted the default route before running each one. So, I was adding a route to a routing table without a 'default' entry each time and displaying the results. I apologize for the confusion. --G

On Sun, Nov 06, 2011 at 09:19:57PM +0000, Keith Williams wrote: I have renumbered around 40 VPSes in the last couple of months, mostly Debian, but some CentOS/Scientific Linux too, and never needed to specify device or metric. But clearly some (all Ubuntu by the looks of it) do. Cheers, Andy -- http://bitfolk.com/ -- No-nonsense VPS hosting

On 6 Nov 2011, at 14:22, Andy Smith &lt;andy(a)bitfolk.com&gt; wrote: I got it too. After some poking around, I ignored it and carried on with the config changes. It all came up nicely after a reboot. James

Hi David, On Mon, Nov 07, 2011 at 09:30:25AM +0200, David Bezuidenhout wrote: You need to remove the default route you have via 212.13.194.1. # ip route del default via 212.13.194.1 It's not causing you any problems right now, but it will when the old IPs go away. Cheers, Andy -- http://bitfolk.com/ -- No-nonsense VPS hosting

On 07/12/2011 15:59, Andy Smith wrote: Following this I've just deleted my old IP address and now Nagios is complaining. The above document says that Nagios will be switched to monitoring the new IP addresses on the 5th :) Regards, Mike

On 07/12/2011 16:20, Andy Smith wrote: Grr, missed that. OK, I'll put the old one back as an alias again. Thanks, Mike

On 07/12/2011 16:28, Andy Smith wrote: Great, many thanks :) Mike

Hi Andy, On 07/02/12 20:36, Andy Maddock wrote: You should still be able to log in using the console service: You will need the password to log into the console. If you haven't changed it then the password is the one you received in the /root/PASSWORDS file that was there when you opened your VPS. If you don't have this then I would email support(a)bitfolk.com and request a password reset. Once the console is open you should be able to log in with root and your root password then follow the instructions at https://tools.bitfolk.com/wiki/Renumbering_for_customers When you're complete, log out with CTRL-D then kill the session with "CTRL-A k". Log back in with the ssh command and password above. Hit ENTER to get the Xen prompt and type "shutdown". Once the VPS is down, type "boot" to bring it back up. You should then be able to reconnect to it using the new IP address. Good luck, Paul.

Hello, On Wed, Feb 08, 2012 at 12:04:04AM +0000, Paul Stimpson wrote: Oh, and the other thing is that there's a password reset link on the login page of https://panel.bitfolk.com/ So as long as your email address still works, you don't need to contact support to ask for a password reset. Cheers, Andy